Evopol: a Framework for Optimising Social Regulation Policies

The aim of the paper is to introduce EvoPol (EVOlving POLicies): an evolutionary computation (EC) technique to optimize the if-then rules to support governmental policy analysis in restricting recruitment of smokers. EC is a population based adaptive method, which may be used to solve optimization problems, based on the genetic processes of biological organisms. Parameters of the fuzzy inference system could be adapted using a neural network learning technique (neuro-fuzzy system) or by evolutionary computation. The fuzzy decision system (FDS) was developed based on three sub-systems: fuzzification, fuzzy knowledge base (if-then rules) and defuzzification. In this paper we propose the fine-tuning of fuzzy rule base and membership function parameters using an evolutionary algorithm. We compare the present work with our previous work using neuro-fuzzy techniques. Empirical results clearly show that evolutionary learning could perform better (improvement of decision score) than the neuro-fuzzy techniques at the expense of extra computation cost.